Mass spectrometry-based approaches are the reference techniques for the determination of nitrite and nitrate in plasma and serum. However, due to their simplicity and rapidity, assays based on the Griess reaction or HPLC are generally used in clinical studies, but they generate diverging values for nitrite/nitrate concentration. In this study, particular attention is paid to the optimization of the deproteinization procedure for plasma and serum samples prior to nitrite/nitrate analysis by an enzymatic batch Griess assay, HPLC and GC-MS. A method is reported to verify completeness of deproteinization and to correct for nonspecific contribution to the absorbance of the diazo dye at 540 nm. With the application of such optimized procedures, we were able to significantly improve the correlation between Griess and HPLC method or the GC-MS technique for nitrite+nitrate concentrations in human serum and plasma. Despite remaining potentially interfering pre-analytical and analytical factors, the procedures reported in the present study may be helpful in a critical evaluation of limits and possibilities of the enzymatic batch Griess assay as a large-scale method for nitrite/nitrate determination in human serum in clinical studies.